Proper Name Retrieval
Running head: PROPER NAME RETRIEVAL
Semantic Category Moderates Phonological Priming of Proper Name Retrieval During Tip-ofthe-Tongue States
Katherine K. White1, Lise Abrams2, and Elizabeth A. Frame1
1Rhodes
2University
College
of Florida
Author Note:
We thank Greg Palm, Rachel Stowe, and Keshav Kukreja for assistance with stimuli
development and data collection, and Meagan Farrell for comments on an earlier draft of this
manuscript. Portions of this research were presented at the 23rd annual convention of the
American Psychological Society, Washington, DC, and the 52nd annual meeting of the
Psychonomic Society in Seattle, WA. Correspondence concerning this article should be sent to
Katherine White, Department of Psychology, Rhodes College, 2000 N. Parkway, Memphis, TN
38112. E-mail: whitek@rhodes.edu.
1
Proper Name Retrieval
2
Abstract
Despite evidence that the majority of tip-of-the-tongue (TOT) states occur for proper names,
little research has investigated factors that influence their resolution. Although phonological
primes typically increase TOT resolution, the present experiment investigated whether priming
effects are mitigated by semantic competition. Participants read questions whose answers were
proper name targets (e.g., Helen Hunt, Elton John) from various semantic categories (e.g., actor,
musician). Following a TOT, another question was presented that either included a prime name
that varied in phonological overlap with the target (full first name or first syllable) and semantic
category (same profession, different profession), or was phonologically and semantically
unrelated to the target. After presenting the target question a second time, participants were more
likely to resolve TOTs following first-name primes than unrelated names, independent of
semantic category. In contrast, first-syllable primes marginally facilitated TOT resolution when
the prime was in a different semantic category but not when the prime was in the same semantic
category. These results demonstrate that semantic overlap increases competition from
phonologically-related names when there is incomplete phonological input, allowing an
alternative name to prevent TOT resolution.
Keywords: tip-of-the-tongue (TOT) states, proper names, phonological priming, semantic
competition
Proper Name Retrieval
3
Semantic Category Moderates Phonological Priming of Proper Name Retrieval
During Tip-of-the-Tongue States
Almost everyone has experienced a time when a famous name evaded retrieval: When
attempting to recall the singer of Candle in the Wind, the name Elton John may be temporarily
unavailable for retrieval. Oftentimes this frustrating memory lapse, known as a tip-of-the-tongue
(TOT) state (e.g., Brown & McNeil, 1966), is accompanied by a persistent and salient name that
one knows is not correct (e.g., Elvis Presley). These alternate names frequently share various
characteristics with the intended "target" name, including phonological features such as first
letter and number of syllables, and syntactic features such as part of speech (e.g., Brown &
McNeill, 1966; Burke, MacKay, Worthley, & Wade, 1991; Miozzo & Caramazza, 1997).
Although not systematically documented, alternates can sometimes share both phonological and
semantic features with the target (e.g., Brown, 1991; Burke et al., 1991), as with musicians Elton
and Elvis whose names also have the same first syllable. Another characteristic of alternates is
that when present during a TOT, they reduce the likelihood of retrieving the target, relative to
TOTs that are not accompanied by alternates (Burke et al., 1991). The present research explored
a possible link between the target and alternate's sharing phonological and semantic features and
a subsequent reduction in target retrieval (i.e., TOT resolution).
Although TOTs are often described as lapses in efficient word retrieval, they are also
particularly useful tools for investigating the processes that underlie speech production. Theories
of speech production generally agree that word retrieval requires selecting a lemma (i.e., a
lexical representation that is semantically and syntactically specified) from all lemmas stored in
one's mental lexicon, followed by phonological encoding of that lemma (e.g., Dell, 1986; Levelt,
1989). TOTs are thought to result from a breakdown between the two processes, lexical selection
Proper Name Retrieval
4
and phonological encoding, such that lexical selection was completed successfully but
phonological encoding was not. Why would lexical selection, but not phonological encoding, be
successful? Multiple connections link semantic representations to corresponding lemmas, and
activation from these converging connections allows selection of the lemma to occur relatively
easily. In contrast, single connections link lemmas to each of their phonological nodes. Whereas
connections to any node can become weak if not used frequently or recently, nodes with single
connections are particularly susceptible to deficits in the transmission of activation because they
do not benefit from a convergence of activation from multiple sources. Thus, TOTs result when
weakened phonological connections do not receive sufficient top-down activation from their
lemma (e.g., James & Burke, 2000; Rastle & Burke, 1996).
Although TOTs occur for all types of words (e.g., nouns, adjectives, verbs), proper names
are particularly susceptible to retrieval deficits (e.g., Burke et al., 1991; Evrard, 2002; Hanley,
2011). As with all word types, retrieval of a proper name requires lemma selection and
phonological encoding. However, within theories of speech production, the dominant theoretical
explanation of increased TOTs for proper names (see Figure 1) proposes an additional singlesource connection that must be traversed before phonological encoding begins (for details see
Burke et al., 1991, and Valentine, Brennen, & Brédart, 1996; but see Bredart, 1993, and Hanley,
2011, for other explanations of why there are more TOTs for people's names relative to nonnames). Proper names such as Elton John require activation of the lemmas for each individual
name (first and last) via single connections from the proper name phrase. Thus, proper names are
especially susceptible to TOTs because retrieving a full name requires activation of both
individual names via single connections, and both names must then be phonologically encoded
(Burke et al., 1991; Burke, Locantore, Austin, & Chae, 2004; Valentine et al., 1996). Consistent
Proper Name Retrieval
5
with this theory, more TOTs occur for three-word names (which have more single-source
connections) than two-word names (Hanley & Chapman, 2008). Furthermore, words presented
as occupations (e.g., a farmer) are easier to learn and recall than identical words presented as
proper names (e.g., Kelly Farmer; James, 2004; McWeeny, Young, Hay, & Ellis, 1987).
Whereas the noun farmer and the proper name phrase Kelly Farmer can both benefit from
multiple converging connections from their semantic nodes to achieve lexical selection, the
additional step of activating the last name lemma Farmer makes its retrieval more difficult than
the occupation.
Phonological priming
Given that TOTs are thought to result from insufficient activation of phonology, previous
research has investigated the influence of phonological primes1 (i.e., words that contain some or
all aspects of the target's phonology) on TOT incidence and resolution. Phonological primes
provide a vehicle by which the weakened lexical-to-phonological connections that cause TOTs
can be strengthened, which can prevent TOTs from occurring in the first place (e.g., James &
Burke, 2000, Experiment 1). With respect to resolving TOTs, phonological priming studies use a
methodology where a TOT is first induced via a general knowledge question. Phonological
primes (or unrelated words as a control) are then presented, followed by the original TOTinducing question. TOT resolution (retrieving the intended target) is measured as a function of
whether primes or unrelated words were presented. Compared to unrelated words, target retrieval
is increased when participants are primed with words that share phonology with the target
(James & Burke, 2000, Experiment 2), specifically the first syllable (e.g., Abrams, White, &
Eitel, 2003; White & Abrams, 2002). Even though TOTs are most common for proper names,
the majority of studies that involve phonological priming have focused on TOTs for non-proper
Proper Name Retrieval
6
names (i.e., low-frequency words from other grammatical classes such as nouns and verbs). One
exception was a study by Burke et al. (2004), who demonstrated that TOTs were less likely to
occur when phonological primes were presented before the target name. Specifically, naming
homophone primes (e.g., cherry pit) both reduced TOT states and increased correct retrieval of
famous target names whose last names were identical to the homophone's phonology (e.g., Brad
Pitt).
All of these findings that phonological primes facilitate word retrieval contradict the
notion that related words can cause TOTs (i.e., the blocking hypothesis; e.g., Jones, 1989; Jones
& Langford, 1987). However, we do know that once in a TOT, alternate words can have a
"blocking" effect by delaying access to a to-be-retrieved target (Burke et al., 1991), suggesting
that there must be some circumstances that increase the difficulty of retrieving a target. In
support of this hypothesis, Abrams and Rodriguez (2005; see also Abrams, Trunk, & Merrill,
2007) demonstrated that a phonological prime only facilitated TOT resolution when the prime
and target differed in part of speech. When a phonological prime shared part of speech with a
target, the degree of TOT resolution was equivalent to resolution following unrelated words,
indicating no priming. Abrams and colleagues interpreted these results in terms of grammatical
class creating competition for lexical selection (e.g., Dell, 1986; MacKay, 1987): The most
active lemma within a particular grammatical class is selected (in this case, the prime), and its
activation level must subside before another word in that grammatical class (the target when it is
the same part of speech) can be selected for production. The constraints imposed by grammatical
class on TOT resolution are exacerbated because the prime and target also share phonology,
making them especially competitive with each other for retrieval.
Semantic priming
Proper Name Retrieval
7
Unlike the facilitation in target retrieval that results from phonological primes, there
appear to be no effects on TOT incidence when participants are primed with solely semantic
primes: Cross and Burke (2004) found that producing a character’s name (e.g., Eliza Doolittle)
did not increase TOTs when later asked to produce the corresponding actor's name (Audrey
Hepburn). Similar evidence has been shown with semantic cues having no influence on TOT
incidence for non-names (Meyer & Bock, 1992). Thus, there appears to be little evidence that
semantic alternates, when presented implicitly through priming or explicitly through cuing,
influence TOTs by blocking their retrieval when no phonological overlap is present.
Theoretically, a semantically-related alternate word can impair retrieval of a target if it creates
competition at lexical selection, which results in slower selection of the correct target, as shown
in picture word interference tasks with semantic distractors (e.g., Schriefers, Meyer, & Levelt,
1990). However, because lexical selection has already been completed when a TOT occurs,
semantic priming should not have an effect on the phonological encoding stage and influence
TOTs, consistent with the above findings.
Interactions between phonology and semantics
Although semantic-only primes have had little influence on TOTs in previous research,
we tested whether semantic overlap between primes and targets could constrain phonological
priming of TOT resolution. In other speech production tasks like picture naming, participants are
asked to name a target picture while ignoring a prime2, but the prime can nonetheless influence
target naming. For example, semantic overlap between primes and targets has been shown to
mediate phonological priming effects on production (e.g., Cutting & Ferreira, 1999; Taylor &
Burke, 2002). In these experiments, production of a target is influenced not by the prime per se
but by a word related to the prime. For example, Cutting and Ferreira (1999) presented primes
Proper Name Retrieval
8
(e.g., prom) that were semantically related to one meaning of a homonym (e.g., ball) and then
asked people to name target pictures that represented the other meaning of the homonym (e.g., a
picture of a toy ball). Relative to unrelated words, primes facilitated picture naming. However,
when there is the potential for competition from a semantically-related word at lexical selection,
mediated priming from phonology is no longer facilitatory. Abdel Rahman and Melinger (2008)
used Dutch primes (e.g., dokter, mandarijn) that were phonologically related to a nonpresented
word that was a semantic competitor (e.g., dolfijn [meaning dolphin]) of the target picture (e.g.,
haai [meaning shark]), and primes produced slower naming times of targets compared to
unrelated words.
If semantic competition can eliminate the facilitatory effects of phonological priming on
picture naming, then it may have a similar influence on TOT resolution. Testing whether
semantic competition constrains phonological priming of TOT resolution would be virtually
impossible to assess with low-frequency non-names (e.g., bandanna) that are often used to
induce TOTs because of the difficulty in finding primes and targets that are both phonologically
and semantically related. However, proper names are ideal primes and targets because they can
overlap phonologically (e.g., by sharing either the first or last name) and semantically (e.g., by
sharing the same profession).
Predictions within theories of speech production
This experiment manipulated the semantic relationship between primes and targets as
well as the amount of phonology shared by primes and targets. Specifically, primes were from
either the same or different profession as the target and either shared partial (i.e., first syllable) or
full (i.e., first name) phonology with the target. For example, Elvis and Elton share first-syllable
phonology and are from the same profession, whereas Helen Keller and Helen Hunt share first
Proper Name Retrieval
9
name phonology and are from a different profession. Two current theories of speech production
were used to derive hypotheses about the influences of these variables on TOT resolution.
Interactive activation theories postulate a bidirectional spread of activation between lemmas and
their phonological representations (e.g., Dell, 1986; MacKay, 1987), whereas discrete theories
(e.g., Levelt, 1989; Roelofs, 2004) maintain feedforward-only activation between lemmas and
phonological nodes. Whereas both theories predict phonological priming of TOT resolution, they
differ in their predictions for whether semantic competition could mediate the effects of
phonological priming.
With respect to phonological priming of TOT resolution, interactive activation theories
(represented in Figure 2) maintain that presentation of a prime (Elmer Fudd) activates its lemma,
which sends top-down activation to its phonological representations (e.g., the syllable node /ɛl/).
Activation then feeds back from phonology to associated lemmas sharing that phonology (e.g.,
Elmo, Elliot), one of which is the target (Elton). This phonological feedback increases the
target’s chance of being resolved, unlike a phonologically-unrelated word which provides no
feedback to the target. Alternatively, discrete theories explain phonological priming by
postulating connections between perceptual and production systems, where phonological primes
activate related lemmas in the perceptual system, which then facilitates production of the target
by activating the target’s phonology in the production system. In the present experiment, we
anticipated that both first name and first syllable primes would facilitate TOT resolution in the
absence of shared semantic features, consistent with speech production theories and with
previous research on non-names (e.g., James & Burke, 2000; Abrams et al., 2003).
However, only interactive activation theories predict an interaction of phonological
priming with semantic category, where phonological primes sharing semantic category with
Proper Name Retrieval
10
targets will compete for retrieval, thus preventing TOT resolution relative to unrelated names
(see right side of Figure 2). During the initial attempt to retrieve the target, all lemmas that
correspond to the semantically-appropriate concept (e.g., musician) will be partially activated,
creating a situation where semantically-related words could compete with the target for retrieval
(e.g., Schriefers, Meyer, & Levelt, 1990). Following a TOT, presentation of a semanticallyrelated prime that is also phonologically related (e.g., Elvis Presley) increases its competitiveness
because the prime receives both top-down activation from semantics and feedback activation
from its phonology. This converging activation to the prime prevents the target (Elton John)
from being activated as long as the prime remains active. Unlike interactive activation theories,
discrete theories do not recognize feedback as a property of the lexical system and therefore
would not predict an interaction between phonological primes and semantic category.
If semantic competition can be created from phonological primes, as predicted by
interactive activation theories, the degree of phonological overlap may also be relevant to the
circumstances under which semantic competition at lexical selection occurs. In contrast to the
above-described predictions for partial phonological primes, presenting the target’s entire first
name phonology (via the prime) may provide the target with sufficient phonological feedback to
activate its lemma, minimizing effects of semantic competition from the prime. Because
feedback activation from phonology to lemmas is thought to be relatively weak (Abdel Rahman
& Melinger, 2008; Dell & O'Seaghdha, 1991, 1992), additional phonological feedback provided
by a full-name prime could be sufficiently strong to enable activation of the target’s lemma
regardless of top-down activation that the prime receives from its semantic nodes. Alternatively,
having the entire first name could increase competition for retrieval when both the prime and
target are in the same semantic category because phonological feedback from all of the first
Proper Name Retrieval
11
name's phonemes is also going to the prime's lemma. A prime that only shares partial phonology
with a target will not receive as much phonological feedback as a full name prime, thus
providing the target a greater chance to reach activation. In sum, primes that share entire first
name phonology with the target are expected to provide more feedback activation to lemmas
relative to primes that share partial phonology. Whether this increased activation will result in
reduced competition (due to the boost in feedback activation to the target's lemmas) or greater
competition (due to sustained activation of the prime) was assessed in the present experiment.
Method
Participants
Participants included 136 Rhodes College undergraduates (92 females, 44 males) who
reported English as their first language. All participants were between the ages of 18 and 22
years (M = 19.16, SD = 1.19) and received course credit or extra credit for participating.
Materials
Ninety famous target names were selected from the following professions3: actors,
comedians, directors, or models (47%); musicians (16%); TV hosts, authors, or journalists (8%);
athletes (9%); politicians or historical figures (8%), and characters (12%). Forty-five target
names were assigned two phonological primes that shared complete first-name phonology with
the target, and 45 target names were assigned two phonological primes that shared the first
syllable with the target's first name.4 Primes differed in their semantic relatedness to the target,
defined by each name's profession; one prime was from the same semantic category as the target
(e.g., both were actors), whereas the other prime was from a different semantic category.
Although each target was assigned a prime that shared or differed in semantic category with the
target, and presentation of primes was counterbalanced across participants, we were unable to
Proper Name Retrieval
12
fully counterbalance phonological overlap. Finding four prime names (two first-name and two
first-syllable primes, one of which shared and one of which differed in the target’s semantic
category) was impossible, so different targets were assigned primes that shared full first-name
phonology or partial first-syllable phonology. Targets and primes always shared gender and were
frequently matched on race (82%). In addition to the two primes assigned to each target, targets
were assigned an unrelated name that did not share phonology or semantic category with its
corresponding target. Half of the unrelated names matched the same semantic category prime's
number of syllables, while the other half matched the different semantic category prime's number
of syllables. When presenting a prime, the computer program randomly selected whether the
prime was in the same category, different category, or was unrelated, but also ensured that each
of these was chosen equally often.
Questions were created for target, prime, and unrelated names. Target questions
described specific semantic information leading to answers that were the target names. In
contrast, questions for the prime and unrelated names embedded the names at the beginning of
the questions and asked some relatively easily retrieved fact about those people. Example targets,
primes, unrelated names, and questions are shown in Table 1. A recognition test was created for
participants to identify target names that they did not retrieve after the second presentation of the
target question. The target question was presented along with four answer choices: the target and
three other names that shared the same semantic category, age group, and gender as the target.
The target was counterbalanced equally often across the four possible answer choice positions (a,
b, c, and d), with one additional (a) and one additional (b) choice. A post-experiment
questionnaire assessed participants’ awareness of the phonological relationship between targets
and primes as well as their intent to strategically use the relationship to resolve TOT states. A
Proper Name Retrieval
13
program written in Visual Basic 5.0 was used to present all stimuli and record participants’
responses.
Design and Procedure
The experimental design was 3 (Prime Condition) x 2 (Target Type) within-subjects
design. Prime Condition had three levels: (1) a phonological prime in the same semantic
category as the target, (2) a phonological prime in a different semantic category from the target,
or (3) a phonologically and semantically unrelated name (see Table 1). Target Type contained
two levels: (1) targets whose phonological primes shared first-name phonology, and (2) targets
whose phonological primes shared first-syllable phonology.
An experimenter provided participants with written and verbal instructions, describing
TOT states as the experience when one is temporarily unable to retrieve a well-known word from
memory. The computer program presented target questions in random order, and after reading
each question participants indicated "known," "unknown," or "TOT." Upon indicating "known",
participants provided the answer and were then presented with a new target question. After
"unknown" or "TOT" responses, participants were presented with either a prime question (same
or different semantic category) or an unrelated question. After answering the prime/unrelated
question, the original target question was presented a second time, and retrieval of the target
name was reattempted, with participants again indicating whether they knew, did not know, or
were still having a TOT. After all 90 target questions were presented, participants completed the
recognition test. Finally, the post-experiment questionnaire was administered.
Results
Initial Responses. Initial "known" and "TOT" responses to target questions were categorized as
either correct or incorrect. Known responses were correct if participants produced the target
Proper Name Retrieval
14
name. TOT responses were considered correct if the target name was produced on the second
presentation of the question or if the correct answer choice was selected on the recognition test
(these responses are often referred to as “positive TOTs”, e.g., Brown & McNeill, 1966). Initial
responses to targets are shown in Table 2, separately for targets that were subsequently primed
with first-name primes vs. first-syllable primes. Despite using different targets in the first-name
and first-syllable prime conditions, the overall patterns of responding were generally similar.
Target Retrieval following Correct "TOT" Responses. In order to assess the influence of primes
that were phonologically and/or semantically related to our intended targets on TOT resolution,
only correct TOTs were included in analyses. A 3 (Prime Condition) x 2 (Target Type) repeatedmeasures analysis of variance (ANOVA) was performed by participants5 on the mean proportion
of target retrieval following the second presentation of the target question (see Figure 3). Fortyseven (35%) participants were unable to be included in this analysis because they did not have at
least one correct TOT in one of the six conditions. A marginal main effect of target type, F(2,
176) = 2.83, MSE = .11, p = .096, η2 = .03, revealed slightly more TOT resolution for targets
assigned to first-name primes (M = 33.5%) than targets assigned to first-syllable primes (M =
28.7%). A significant main effect of prime condition, F(2, 176) = 6.33, MSE = .11, p = .002, η2 =
.07, showed priming of TOT resolution such that resolution following phonological primes in a
different semantic category (M = 38.0%) was greater than resolution following unrelated names
(M = 25.7%), p < .001, η2 = .11, whereas resolution following phonological primes in the same
semantic category (M = 29.5%) was equivalent to resolution following unrelated names, p =
.284, η2 = .01. These effects were moderated by a significant Prime Condition x Target Type
interaction, F(2, 176) = 3.12, MSE = .10, p = .047. For targets assigned to first-name primes,
significant priming of TOT resolution occurred following both same- and different-category
Proper Name Retrieval
15
phonological primes: relative to unrelated names, more TOT resolution occurred following
same-category primes (p = .015, η2 = .07) as well as different-category primes (p = .005, η2 =
.09). Similar resolution was found following same- and different-category primes (p = .534, η2 =
.004). For targets assigned to first-syllable primes, marginally significant priming of TOT
resolution occurred following different-category primes, which led to greater resolution than
unrelated names (p = .072, η2 = .04) and same-category primes (p = .004, η2 = .09). However,
same-category primes did not show priming, resulting in similar TOT resolution as unrelated
names, p = .314, η2 = .01.
The post-experiment questionnaire was used to categorize participants as being aware of
the priming manipulation and, more importantly, intending to use primes to help retrieval of
targets. Sixty-one participants in the above analysis were categorized as aware of the relationship
between the primes and targets because they mentioned the phonological relationship between
the primes and targets. Despite this awareness, relatively few of these participants (N=25) also
demonstrated intent to use the primes as a way to resolve targets. To ensure that strategic
retrieval processes were not the cause of the priming results, only the 64 participants categorized
as not having intent were included in a 3 (Prime Condition) x 2 (Target Type) ANOVA.
Although there was no main effect of target type, F < 1, there was a significant main effect of
prime condition, F(2, 126) = 5.45, MSE = .11, p = .005, η2 = .06, and a marginal Prime Condition
x Target Type interaction, F(2, 126) = 2.76, MSE = .10, p = .068. Following up this interaction
revealed the same pattern of results as the analysis including all participants.
Discussion
This experiment demonstrated a novel constraint on TOT resolution: Shared semantic
category limited the ability of a prime's phonology to increase TOT resolution, but only when the
Proper Name Retrieval
16
prime did not fully overlap in phonology with the target’s first name. Furthermore, these effects
did not result from participants using strategic processes to elicit the target's retrieval. The lack
of priming of TOT resolution when primes and targets shared semantic category and only partial
(first syllable) phonology is explained by theories that assume feedback from phonological nodes
to associated lemmas (e.g., Dell, 1986; MacKay, 1987). The phonological feedback, coupled
with top-down semantic activation to the prime, is sufficient to maintain activation of the prime,
which results in delayed target retrieval for as long as the prime remains active (see Figure 2).
Although phonological feedback is also sent to the target, feedback from only one syllable is
relatively weak, allowing for potential competition from semantically-related words (e.g., Abdel
Rahman & Melinger, 2008). This competition suggests that constraints on lemma activation
during TOT resolution for names can come from semantics, similar to findings from speech
production studies with non-names (e.g., Schriefers et al., 1990). In contrast to first-syllable
primes that shared semantic features with the target, when a prime shared entire first name with a
target, it facilitated TOT resolution. This finding suggests that sufficient phonological overlap
can offset potential semantic competition and provide enough activation to the target’s lemma to
boost TOT resolution.
In the absence of semantic overlap, both first-name and first-syllable phonological primes
increased TOT resolution, consistent with both interactive activation (e.g., Dell, 1986) and
discrete (e.g., Levelt, 1989) theories of speech production. These results emphasize the
importance of the first syllable in resolving TOTs for proper names, again paralleling findings of
priming with non-names (e.g., Abrams et al., 2003; White & Abrams, 2002). However, the
priming effect following first-syllable primes was marginal (M = 9.6), compared to a highly
significant priming effect (M = 12.1) following first-name primes. While we must interpret these
Proper Name Retrieval
17
results cautiously since different target names are being compared, these results suggest the
extent of phonological priming may be contingent on the amount of phonological feedback, as
first-name primes provide more phonological feedback than first-syllable primes. As illustrated
in Figure 2, phonological priming of names that share first syllable would occur when feedback
travels from the first syllable /ɛl/ to Elton and then across an additional connection to the proper
name phrase Elton John. In order to retrieve the target, bottom-up activation must be sufficient to
facilitate activation of the remainder of Elton John’s phonology. However, less priming might be
expected from first-syllable primes than from first-name primes due to insufficient bottom-up
activation being transmitted from a single syllable. Another explanation for the reduction in
phonological priming is that in the present task, two words (both names) must get activated, and
the ability of a single syllable to facilitate retrieval of both names may become increasingly more
difficult.
Demonstrating that semantics and phonology interact to influence TOT resolution offers
unique insight into the underlying processes of lexical selection and phonological encoding.
TOTs are special cases where production is unsuccessful and indeed is momentarily "frozen"
between two stages, lexical selection and phonological encoding. Thus, whereas the picture
naming task provides insight into processes that affect (successful) lemma selection, the TOT
resolution methodology provides insight into processes that occur at a later stage in production,
namely after a lemma has been selected but before it has been successfully phonologically
encoded. The role of semantic competition has clearly been demonstrated prior to lexical
selection via picture naming (e.g., Schriefers et al., 1990), but the present study has shown that
semantics continues to play a role in the later stages of production (i.e., after lexical selection has
occurred) by interfacing with phonology to constrain which associated lemmas get activated. The
Proper Name Retrieval
18
interactive effect of semantics and phonology on TOT resolution also parallels previous studies
using other speech production tasks, such as picture naming (Abdel Rahman & Melinger, 2008;
Cutting & Ferreira, 1999; Taylor & Burke, 2002) as well as studies of speech errors for proper
names (Brédart & Valentine, 1992; Martin, Weisberg, & Saffran, 1989), which find a greater
proportion of mixed phonological and semantic substitutions compared to solely semantic or
phonological errors.
Although the comparison of names and non-names was not manipulated in the present
experiment, using TOT resolution to investigate proper name retrieval may also provide insight
into why proper names are particularly difficult to learn and retrieve. One way in which proper
names differ from non-names is that the latter are more likely to have synonyms (e.g., movie,
film, show) that can be substituted during speech production in order to not disrupt the flow of
speech (Cohen & Faulkner, 1986). Evidence supporting this view was recently shown by Hanley
(2011), who found no differences in proper and common name retrieval when controlling for the
number of alternative words or names that could be used in place of a target. Further, Brédart
(1993) found more correct naming of faces that had two names (a real name and a character
name) associated with them than faces that had one name associated with them. Another
potential source for the increased difficulty in retrieving proper names is the frequency with
which the name is encountered (e.g., Hanley & Chapman, 2008), supported by evidence that
more common names (e.g., Davis) are easier to learn and recall than less common names (e.g.,
Davin; James & Fogler, 2007). The results of the present study offer additional conditions that
influence why proper names are oftentimes more difficult to retrieve than non-names: Proper
names may be more likely to have semantic competition along with phonological overlap,
conditions that impeded TOT resolution in our study. Although some non-names are
Proper Name Retrieval
19
semantically- and phonologically-related (e.g., chastity and charity), they are certainly more rare
than proper names that can share semantic features such as profession.
Our data also offer behavioral evidence consistent with recent claims that proper and
common names are represented differently at the neural level (Semenza, 2006, 2009). The
present experiment found phonological priming between names without semantic overlap,
suggesting that proper names are not competitive with each other, independent of other shared
features. In contrast to proper names, phonologically-similar non-names are quite competitive, as
illustrated by a lack of phonological priming when primes and targets share grammatical class
(Abrams & Rodriguez, 2005; Abrams et al., 2007). Thus, proper names seem to function
differently from non-names in terms of the role that grammatical class plays in their production.
In conclusion, the present results suggest one potential reason for why alternates might
delay TOT resolution in everyday life: Alternates can share both phonology and semantic
features with targets, causing them to compete for retrieval. It is worth noting that inhibitory
effects on TOT resolution were not observed in the present experiment (i.e., where TOT
resolution following unrelated names was greater than same-category primes). However, it is
possible that inhibitory effects from shared semantic category would be easier to detect in old
age (see Abrams et al., 2007), given that TOTs increase with age (Burke et al., 1991), especially
for proper names (e.g., Evrard, 2002; Rendell, Castel, & Craik, 2005). Older adults' enhanced
semantic networks, combined with age-related deficits in the transmission of phonological
activation, could potentially make them more susceptible to competition from phonologicallyand semantically-related alternate names and result in a reduction in TOT resolution. More
generally, continued investigations of the circumstances under which TOT resolution for proper
names is either facilitated or inhibited will lead to a better understanding of the interactions
Proper Name Retrieval
20
between lexical selection and phonological encoding. These stages of word retrieval are critical
for distinguishing between discrete and interactive activation theories of speech production and
will give further insight into why names are particularly vulnerable to retrieval deficits.
Proper Name Retrieval
21
References
Abdel Rahman, R., & Melinger, A. (2008). Enhanced phonological facilitation and traces of
concurrent word form activation in speech production: An object-naming study with
multiple distractors. The Quarterly Journal of Experimental Psychology, 61, 1410–1440.
doi: 10.1080/17470210701560724.
Abrams, L., & Rodriguez, E. L. (2005). Syntactic class influences phonological priming of tipof-the-tongue resolution. Psychonomic Bulletin and Review, 12, 1018–1023.
Abrams, L., Trunk, D. L., & Merrill, L. A. (2007). Why a superman cannot help a tsunami:
Activation of grammatical class influences resolution of young and older adults’ tip-ofthe-tongue states. Psychology and Aging, 22, 835–845. doi: 10.1037/0882-7974.22.4.835
Abrams, L., White, K. K., & Eitel, S. L. (2003). Isolating phonological components that increase
tip-of-the-tongue resolution. Memory and Cognition, 31, 1153–1162.
Baayen, R. H., Tweedie, F. J., & Schreuder, R. (2002). The subjects as a simple random effect
fallacy: Subject variability and morphological family effects in the mental lexicon. Brain
and Language, 81, 55–65. doi: 10.1006/brln.2001.2506
Brédart, S., & Valentine, T. (1992). From Monroe to Moreau: An analysis of face naming errors.
Cognition, 45, 187–223.
Brown, A. S. (1991). A review of the tip-of-the-tongue experience. Psychological Bulletin, 109,
204–223.
Brown, R. & McNeill, D. (1966). The "tip of tongue" phenomenon. Journal of Verbal Learning
and Verbal Behavior, 5, 325–337. doi: 10.1016/S0022-5371(66)80040-3
Proper Name Retrieval
22
Burke, D. M., Locantore, J. K., Austin, A. A., & Chae, B. (2004). Cherry pit primes Brad Pitt:
Homophone priming effects on young and older adults' production of proper names.
Psychological Science, 15, 164–170. doi: 10.1111/j.0956-7976.2004.01503004.
Burke, D. M., MacKay, D. G., Worthley, J. S. & Wade, E. (1991). On the tip of the tongue:
What causes word finding failures in young and older adults? Journal of Memory and
Language, 30, 542–579. doi:10.1016/0749-596X(91)90026-G
Burton, A.M. & Bruce, V. (1992). I recognise your face, but I can't remember your name: a
simple explanation? British Journal of Psychology, 83, 45–60.
Clark, H. H. (1973). The language-as-fixed-effect fallacy: A critique of language statistics in
psychological research. Journal of Verbal Learning and Verbal Behavior, 12, 335–359.
doi: 10.1016/S0022-5371(73)80014-3
Cohen, G., & Faulkner, D. (1986). Memory for proper names: Age differences in retrieval.
British Journal of Developmental Psychology, 4, 187–197.
Cutting, J. C., & Ferreira, V. S. (1999). Semantic and phonological information flow in the
production lexicon. Journal of Experimental Psychology: Learning, Memory, and
Cognition, 25, 318–344. doi: 10.1037/0278-7393.25.2.318
Dell, G. S. (1986). A spreading-activation theory of retrieval in sentence production.
Psychological Review, 93, 283–321. doi:10.1037/0033-295X.93.3.283
Dell, G. S., & O’Seaghdha, P. G. (1991). Mediated and convergent lexical priming in language
production: A comment on Levelt et al. (1991). Psychological Review, 98, 604–614. doi:
10.1037/0033-295X.98.4.604
Dell, G. S. & O'Seaghdha, P. G. (1992). Stages of lexical access in language production.
Cognition, 42, 287–314. doi: 10.1016/0010-0277(92)90046-K
Proper Name Retrieval
23
Evrard, M. (2002). Ageing and lexical access to common and proper names in picture naming.
Brain and Language, 81, 174–179. doi: 10.1006/brln.2001.2515
Hanley, J. R. (2011). Why are names of people associated with so many phonological retrieval
failures? Psychonomic Bulletin & Review, 18, 612–617. doi: 10.3758/s13423-011-00820
Hanley, J. R., & Chapman, E. (2008). Partial knowledge in a tip-of-the-tongue state about twoand three-word proper names. Psychonomic Bulletin and Review, 15, 156- 160. doi:
10.3758/PBR.15.1.156
James, L. E. (2004). Meeting Mr. Farmer versus meeting a farmer: Specific effects of aging on
learning proper names. Psychology and Aging, 19, 515-522. doi: 10.1037/08827974.19.3.515
James, L. E. & Burke, D. M. (2000). Phonological priming effects on word retrieval and tip-ofthe-tongue experiences in young and older adults. Journal of Experimental Psychology:
Memory, Learning, and Cognition, 26, 1378–1391. doi:10.1037//0278-7393.26.6.1378
James, L. E., & Fogler, K. A. (2007). Meeting Mr. Davis versus Mr. Davin: Effects of name
frequency on learning proper names in young and older adults. Memory, 15, 366-374.
doi:10.1080/09658210701307077
Jones, G. V. (1989). Back to Woodworth: Role of interlopers in the tip-of-the-tongue
phenomenon. Memory and Cognition, 17, 69-76.
Jones, G.V., & Langford, S. L. (1987). Phonological blocking in the tip of the tongue state.
Cognition, 26, 115-122.
Levelt, W. J. M. (1989). Speaking: From intention to articulation. Cambridge, MA: MIT Press.
Proper Name Retrieval
24
MacKay, D. G. (1987). The organization of perception and action: A theory for language and
other cognitive skills. New York: Springer-Verlag
McWeeny, K. H., Young, A. W., Hay, D. C., & Ellis, A. W. (1987). Putting names to faces.
British Journal of Psychology, 78, 143–149.
Miozzo, M. & Caramazza, A. (1997). Retrieval of lexical-syntactic features in tip-of-the tongue
states. Journal of Experimental Psychology: Learning, Memory, and Cognition, 23,
1410–1423. doi: 10.1037/0278-7393.23.6.1410
Raaijmakers, J. G. W., Schrijnemakers, J. M. C., & Gremmen, F. (1999). How to deal with "The
language-as-fixed-effect fallacy"; Common misconceptions and alternative solutions.
Journal of Memory and Language, 41, 416–426. doi: 10.1006/jmla.1999.2650
Rastle, K. G., & Burke, D. M. (1996). Priming the tip of the tongue: Effects of prior processing
on word retrieval in young and older adults. Journal of Memory and Language, 35, 585–
605. doi: 10.1006/jmla.1996.0031
Rendell, P. G., Castel, A. D., & Craik, F. I. M. (2005). Memory for proper names in old age: A
disproportional impairment? The Quarterly Journal of Experimental Psychology, 58A,
54–71. doi:10.1080/02724980443000188
Roelofs, A. (2004). Error biases in spoken word planning and monitoring by aphasic and
nonaphasic speakers: Comment on Rapp and Goldrick (2000). Psychological Review,
111, 561–572. doi: 10.1037/0033-295X.111.2.561
Schriefers, H., Meyer, A. S., & Levelt, W. J. (1990). Exploring the time course of lexical access
in language production: Picture-word interference studies. Journal of Memory and
Language, 29, 86–102. doi: 10.1016/0749-596X(90)90011-N
Proper Name Retrieval
25
Semenza, C. (2006). Retrieval pathways for common and proper names. Cortex, 42, 884–91. doi:
10.1016/S0010-9452(08)70432-5
Semenza, C. (2009). The neuropsychology of proper names. Mind & Language, 24, 347-469.
doi: 10.1111/j.1468-0017.2009.01366.x
Taylor, J. K., & Burke, D. M. (2002). Asymmetric aging effects on semantic and phonological
processes: Naming in the picture-word interference task. Psychology and Aging, 17, 662–
676. doi: 10.1037/0882-7974.17.4.662
Valentine, T., Brennen, T., & Brédart, S. (1996). The cognitive psychology of proper names: On
the importance of being Ernest. London: Routledge.
Vitkovitch, M., Potton, A., Bakogianni, C., & Kinch, L. (2006). Will Julia Roberts harm Nicole
Kidman? Semantic priming effects during face naming. The Quarterly Journal of
Experimental Psychology, 59, 1134-1152. doi:10.1080/02724980543000178
White, K. K., & Abrams, L. (2002). Does priming specific syllables during tip-of-the-tongue
states facilitate word retrieval in older adults? Psychology and Aging, 17, 226–235. doi:
10.1037/0882-7974.17.2.226
Proper Name Retrieval
26
Footnotes
1
Using primes to strengthen connections to phonology provides an understanding of TOT
resolution independent of strategic processing, which occurs in cueing paradigms where one can
use a given word to help trigger recall of the target. Although there is a substantial literature on
TOT incidence and resolution when strategic retrieval processes are engaged via presentation of
cues (e.g., Brennen, Baguley, Bright, & Bruce, 1990; Freedman & Landauer, 1966; Heine, Ober,
& Shenaut, 1999; Jones, 1989; Meyer & Bock, 1992; Perfect & Hanley, 1992), we do not discuss
that literature because priming effects are thought to better resemble normal word retrieval
(Meyer & Bock, 1992) and the spontaneous TOT resolution that occurs when phonology is
encountered “inadvertently in conversation” (James & Burke, 2000, p. 1379).
2
Picture naming studies typically refer to these primes as "distractors" because they are intended
to be ignored while participants name pictures. However, for consistency and because phonology
typically facilitates naming in TOT resolution, we use the term "primes".
3
There were two exceptions to the "professions" categorization: historical figures and characters.
Proper names were categorized as historical figures if they had a significant impact on United
States or world history (e.g., Alexander Hamilton, John Locke) and as characters if they were
fictional figures portrayed in movies, television shows, or books (e.g., Peter Parker, Charlie
Bucket).
4
Twenty primes (22%) shared phonology beyond just first syllable with the target (e.g., an
additional phoneme), and these primes were divided almost equivalently across same and
different semantic category conditions. Analysis excluding these primes revealed the same
pattern of results as the analysis including all primes.
Proper Name Retrieval
5
27
Consistent with most studies of TOT resolution, we did not conduct item analyses for several
reasons. TOT rates are typically very low and it cannot be anticipated prior to the experiment on
which items participants will have a TOT, which violates underlying statistical assumptions (e.g.,
Baayen, Tweedie, & Schreuder, 2002; Clark, 1973; Raaijmakers, Schrijnemakers, & Gremmen,
1999). As a result, participants have TOTs on different items, and it is impossible to
systematically select the priming condition for an item, which results in many items being
excluded from analysis, and thus item analyses are less powerful to detect priming effects. In any
case, because participants have TOTs on a different subset of items, the variability across
participants in the participant analyses includes item variability to some extent.
Proper Name Retrieval
28
Table 1.
Example questions, target names, prime names, and unrelated names.
Target Question
Target
Name
Phonological
Prime in
Same
Semantic
Category
Helen Mirren
(actor)
Phonological
Prime in
Different
Semantic
Category
Helen Keller
(historical
figure)
Unrelated
Name
Example Prime/Unrelated
Question
What is the name of the 47
year old blonde female actor
who starred in the movies As
Good As It Gets, Twister, Cast
Away, What Women Want, and
Pay It Forward?
Helen
Hunt
(actor)
What is the name of the retired
football quarterback who
graduated from Notre Dame,
won four Super Bowls with the
San Francisco 49ers between
1979 and 1994, and currently
endorses Sketchers Shape
Ups?
Joe
Namath
(athlete)
Joe Montana
(athlete)
Joe Biden
(politician)
Heath
Ledger
(actor)
Joe Biden, the current Vice
President of the United States, was
previously the senator for what
U.S. state?
Delaware
Alfred
Hitchcock
(director)
Alec
Baldwin
(director/
actor)
Albert
Einstein
(historical
figure)
Herman
Melville
(author)
Herman Melville is the American
author who wrote the classic book
'Moby Dick', which is about what
type of marine mammal?
Whale
Elton John
(musician)
Elvis Presley
(musician)
Elmer Fudd
(character)
Chevy
Chase
(actor)
Elvis Presley, the iconic musician
who sang the songs 'Jailhouse
Rock,' and 'Hound Dog,' owned a
mansion in Memphis, TN that is
called what?
Graceland
What is the name of the man
who directed the old
Hollywood movies 'Psycho,'
'To Catch a Thief,' 'Rear
Window,' and 'The Birds'?
What is the name of the
flamboyant English musician
who sings the songs, 'Tiny
Dancer,' 'Bennie and the Jets,'
and 'Candle in the Wind'?
Martha
Stewart
(business
woman)
Helen Mirren, the 65 year old
British female actor, won an
Academy Award for her leading
role in what 2006 movie?
Answer to
Prime
Question
The
Queen
Proper Name Retrieval
Note: Primes in the first-syllable phonology condition have shared phonology underlined.
29
Proper Name Retrieval
30
Table 2.
Percent (%) of initial responses to targets in first-name and first-syllable conditions.
_____________________________________________
First-Name
First-Syllable
Targets
Targets
_____________________________________________
Correct “Know”
26.9
29.0
Incorrect “Know”
6.4
4.6
Correct “TOT”
16.6
12.2
Incorrect “TOT”
2.4
1.7
Unknown
47.7
52.4
_____________________________________________
Note. Means are based on the 89 participants who were included in the TOT resolution analysis.
Proper Name Retrieval
reddishmusician
British hair
Sings “Candle
in the Wind”
31
SEMANTIC SYSTEM
Propositional Nodes
ELTON JOHN
Lexical Nodes
(lemma)
(proper name phrase)
ELTON
JOHN
(initial name)
(last name)
PHONOLOGICAL SYSTEM
/tn/
/ɛl/
/dʒɒn/
Syllable Nodes
Phonological Nodes
/ɛ/
/l/
/t/
/n/ /dʒ/
/ɒ/
/n/
Figure 1. An illustration of semantic, lexical, and phonological nodes for the proper name Elton
John. Weakened single connections between nodes (e.g., the proper name phrase Elton John and
the initial name Elton and last name John) are represented by dotted lines.
Proper Name Retrieval
32
SEMANTIC SYSTEM
musician
cartoon
character
musician
ELTON JOHN
ELMER FUDD
ELTON JOHN
ELVIS PRESLEY
(proper name
phrase)
(proper name
phrase)
(proper name
phrase)
(proper name
phrase)
ELTON
ELMER
ELTON
Propositional
Nodes
Lexical Nodes
(lemma)
ELVIS
PHONOLOGICAL
SYSTEM
/ɛl/
/tn/
/mər/
/ɛl/
/tn/
/vɪs/
Syllable
Nodes
Figure 2. A simplified illustration of how phonological primes Elmer Fudd and Elvis Presley
affect TOT resolution of Elton John. The left side of the figure illustrates predictions within
interactive activation theories regarding TOT resolution following presentation of a phonological
prime from a semantic category that is different from the target. The right side of the figure
illustrates predictions when presented with a phonological prime from the same semantic
category as the target. Although phonological feedback is simplified to illustrate first syllable,
note that all aspects of phonology (e.g., last syllable) provide feedback.
Proper Name Retrieval
Same Semantic Category
50
Different Semantic Category
45
Unrelated Name
40
% TOT Resolution
33
35
30
25
20
15
10
5
0
First Name Prime
First Syllable Prime
Target Type
Figure 3. Percent target retrieval (TOT resolution) as a function of prime condition
(phonological prime in the same semantic category as the target, phonological prime in a
different semantic category, unrelated name) and target type (targets whose phonological primes
shared first-name phonology, targets whose phonological primes shared first-syllable
phonology).